From mboxrd@z Thu Jan 1 00:00:00 1970 From: alexdeucher@gmail.com Subject: [PATCH 021/165] drm/radeon: implement async vm_flush for the sDMA (v6) Date: Wed, 26 Jun 2013 09:21:41 -0400 Message-ID: <1372253045-17042-22-git-send-email-alexdeucher@gmail.com> References: <1372253045-17042-1-git-send-email-alexdeucher@gmail.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: Received: from mail-qa0-f41.google.com (mail-qa0-f41.google.com [209.85.216.41]) by gabe.freedesktop.org (Postfix) with ESMTP id 50AE2E6351 for ; Wed, 26 Jun 2013 06:24:35 -0700 (PDT) Received: by mail-qa0-f41.google.com with SMTP id f14so1569767qak.0 for ; Wed, 26 Jun 2013 06:24:34 -0700 (PDT) In-Reply-To: <1372253045-17042-1-git-send-email-alexdeucher@gmail.com> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: dri-devel-bounces+sf-dri-devel=m.gmane.org@lists.freedesktop.org Errors-To: dri-devel-bounces+sf-dri-devel=m.gmane.org@lists.freedesktop.org To: dri-devel@lists.freedesktop.org Cc: Alex Deucher List-Id: dri-devel@lists.freedesktop.org From: Alex Deucher Update the page table base address and flush the VM TLB using the sDMA. V2: update for 2 level PTs V3: update vm flush V4: update SH_MEM* regs V5: switch back to old style VM TLB invalidate V6: fix packet formatting Signed-off-by: Alex Deucher --- drivers/gpu/drm/radeon/cik.c | 70 ++++++++++++++++++++++++++++++++++++++++++ 1 files changed, 70 insertions(+), 0 deletions(-) diff --git a/drivers/gpu/drm/radeon/cik.c b/drivers/gpu/drm/radeon/cik.c index 931169e..3c18a63 100644 --- a/drivers/gpu/drm/radeon/cik.c +++ b/drivers/gpu/drm/radeon/cik.c @@ -3407,6 +3407,76 @@ void cik_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm) radeon_ring_write(ring, 0x0); } +/** + * cik_dma_vm_flush - cik vm flush using sDMA + * + * @rdev: radeon_device pointer + * + * Update the page table base and flush the VM TLB + * using sDMA (CIK). + */ +void cik_dma_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm) +{ + struct radeon_ring *ring = &rdev->ring[ridx]; + u32 extra_bits = (SDMA_POLL_REG_MEM_EXTRA_OP(1) | + SDMA_POLL_REG_MEM_EXTRA_FUNC(3)); /* == */ + u32 ref_and_mask; + + if (vm == NULL) + return; + + if (ridx == R600_RING_TYPE_DMA_INDEX) + ref_and_mask = SDMA0; + else + ref_and_mask = SDMA1; + + radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_SRBM_WRITE, 0, 0xf000)); + if (vm->id < 8) { + radeon_ring_write(ring, (VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (vm->id << 2)) >> 2); + } else { + radeon_ring_write(ring, (VM_CONTEXT8_PAGE_TABLE_BASE_ADDR + ((vm->id - 8) << 2)) >> 2); + } + radeon_ring_write(ring, vm->pd_gpu_addr >> 12); + + /* update SH_MEM_* regs */ + radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_SRBM_WRITE, 0, 0xf000)); + radeon_ring_write(ring, SRBM_GFX_CNTL >> 2); + radeon_ring_write(ring, VMID(vm->id)); + + radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_SRBM_WRITE, 0, 0xf000)); + radeon_ring_write(ring, SH_MEM_BASES >> 2); + radeon_ring_write(ring, 0); + + radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_SRBM_WRITE, 0, 0xf000)); + radeon_ring_write(ring, SH_MEM_CONFIG >> 2); + radeon_ring_write(ring, 0); + + radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_SRBM_WRITE, 0, 0xf000)); + radeon_ring_write(ring, SH_MEM_APE1_BASE >> 2); + radeon_ring_write(ring, 1); + + radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_SRBM_WRITE, 0, 0xf000)); + radeon_ring_write(ring, SH_MEM_APE1_LIMIT >> 2); + radeon_ring_write(ring, 0); + + radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_SRBM_WRITE, 0, 0xf000)); + radeon_ring_write(ring, SRBM_GFX_CNTL >> 2); + radeon_ring_write(ring, VMID(0)); + + /* flush HDP */ + radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_POLL_REG_MEM, 0, extra_bits)); + radeon_ring_write(ring, GPU_HDP_FLUSH_DONE); + radeon_ring_write(ring, GPU_HDP_FLUSH_REQ); + radeon_ring_write(ring, ref_and_mask); /* REFERENCE */ + radeon_ring_write(ring, ref_and_mask); /* MASK */ + radeon_ring_write(ring, (4 << 16) | 10); /* RETRY_COUNT, POLL_INTERVAL */ + + /* flush TLB */ + radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_SRBM_WRITE, 0, 0xf000)); + radeon_ring_write(ring, VM_INVALIDATE_REQUEST >> 2); + radeon_ring_write(ring, 1 << vm->id); +} + /* * RLC * The RLC is a multi-purpose microengine that handles a -- 1.7.7.5